Re: Question, Flaky tests: pyspark.sql.tests.ArrowTests tests in Jenkins worker 5(?)

ok, first test to run post-fix is green: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/80289/ i'll keep an eye on this worker over the next few days. shane On Sat, Aug 5, 2017 at 11:06 AM, shane knapp wrote: > amp-jenkins-worker-05 had 0.20.3 installed

Re: Question, Flaky tests: pyspark.sql.tests.ArrowTests tests in Jenkins worker 5(?)

amp-jenkins-worker-05 had 0.20.3 installed for some reason. it's now been downgraded to 0.19.2 and matches the other workers. shane On Sat, Aug 5, 2017 at 2:01 AM, Liang-Chi Hsieh wrote: > > Maybe a possible fix: >

Re: the uniqueSource in StreamExecution, where is it be changed please?

Not sure if you are looking for how the returned value of `getOffset` changes. I think it depends on how the actual `Source` classes implement it. For example, in `FileStreamSource`, you can see `getOffset` is updated by finding new files in the source. Different source has different way to get

Re: Question, Flaky tests: pyspark.sql.tests.ArrowTests tests in Jenkins worker 5(?)

Maybe a possible fix: https://stackoverflow.com/questions/31495657/development-build-of-pandas-giving-importerror-c-extension-hashtable-not-bui Hyukjin Kwon wrote > Hi all, > > I am seeing flaky Python tests time to time and if I am not mistaken > mostly > in amp-jenkins-worker-05: > > >

Question, Flaky tests: pyspark.sql.tests.ArrowTests tests in Jenkins worker 5(?)

Hi all, I am seeing flaky Python tests time to time and if I am not mistaken mostly in amp-jenkins-worker-05: == ERROR: test_filtered_frame (pyspark.sql.tests.ArrowTests)